Tajfel et al 1971

Aim

To test whether the simple act of grouping was enough to produce prejudice between groups of very similar people even when there is no history or competition between the groups.

Sample

An opportunity sample of 64 14-15 year old boys from a Bristol comprehensive school was used, who supposedly already had a cohesive (one) group identity.

Procedure experiment 1

1) tested in a laboratory in eight separate groups of eight boys.

2) taken to a lecture room and told that the study was investigating visual judgements.

3) shown 40 different dot clusters on a screen and asked to estimate the number of dots in each cluster.

4) then divided into two experimental groups. Condition 1 - over-estimators

Condition 2 more accurateIn

5)They were then told they were required to take part in further investigation on other kinds of decision-making.

6) assigned to groups according to the judgements they had made about the number of dots.

7) They were randomly allocated to groups and in condition 1, boys were designated as 'overestimators' or 'underestimators', in condition 2 they were split into 'better' or 'worse' accuracy groups.

8) They were told that the task involved money in rewarding and punishing the other boys and they had to choose how much to reward or punish the other boy, whether they were in their own group or the other group, or whether one was in his group and the other was not.

9) They were taken to separate cubicles and in each cubicle there was a booklet with a matrix which had two rows each with 14 numbers, with each number in a box.

10) The numbers represented the amount of money they were either rewarding or taking from the other boys.

Matrices

There were three kinds of matrices; in-group choices is when the boys in both rows were members of the participants own group, out-group choices is when the boys in both rows were members of the other group, and intergroup choices is when one row identified a member of the participants own group and the other row was from the out-group. One row referred to one boy and one to the other boy, so they had to choose which pair they wanted that would affect both of them. If they gave as much as possible to one boy, this was given a score of 14 and a score of 1 if they had given a boy as little as possible. A fair score would be 7 because that would mean they had allocated rewards/punishments fairly.

Results for experiment 1

It seemed that when decisions involved boys in the same group, the choices were fairer than when decisions were made about 2 boys in different groups.

When decisions involved boys from different groups (in-group/out-group), the results were closely distributed around the point of fairness, with the average score being 7.5. For intergroup decisions (in-group/in-group or out-group/out-group) Tajfel found that the large majority of participants gave more money to members of their in-group, with the average score being 9/14.

Conclusions for experiment 1

They boys new each other well. Groups were defined by 'flimsy and unimportant criteria'. The boy's interests were not much affected since no one knew what their actual choices were. The amount of profit was not trivial- each left with the equivalent of about $1. They could have gone for maximum joint profit (all the boys would end up with the most money), or choose the point of maximum fairness. They did choose maximum fairness when their choices were inter-group ones. In order to analyse their choices further a second study was conducted.

Experiment 2: Klee and Kandinsky study Aim

To investigate factors that lead to in-group favouritism.

Procedure for experiment 2

1) 3 new groups with 16 boys in each group.

2) They were shown 12 slides of 6 paintings by Klee and 6 by Kandinsky and were randomly allocated into two groups on basis of which artist they said they preferred, half were assigned to either the Klee preference group or the Kandinsky group.

3) They were then given a rewards allocation task where they were asked to award points to two other boys (one from each group) at a time. They only knew what group they were in.

4) The matrices they had to use were arranged so that when a participant chose a score for one boy, the other boy would get the score that was tied to it.

5) There were 3 conditions in the way the boys could allocate money to two other boys, either to two boys from their own group (in-group), two boys from the other group (outer-group) or one boy from their group and one boy from the other group (inter-group).

6) The researchers were interested whether the boys would choose the matrix that gave both groups the greatest possible reward (maximum joint profit), the matrix that gave the in-group the greatest reward regardless of what this meant for the out-group (maximum in-group profit) or the matrix that created the biggest difference between the in-group and the out-group (maximum difference.)

Results for experiment 2

It was found that the boys typically awarded more points to members of their in-group even though this meant they ended up with less money themselves.

Maximum in-group profit and maximum difference in favour of the in-group worked against maximum joint profit. If they had a choice between giving the most money to everyone and giving the most money to their own group, they favoured their own group. Even if giving more to the othergroup did not mean giving less to their own group, they still gave more to their own.

Conclusions for experiment 2

Categorising the boys into meaningless groups caused them to identify with their in-group and generate a positive social identity by giving their group more points. They did this even if it meant ending up with fewer points, as long as their own group came out on top. This in-group favouritsm can be explained by the social identity theory (proposed by Tajfel, 1970) which suggests that the boys made themselves superior by boosting their group's status by awarding themselves more money. The theory suggests that to boost ones own self-esteem, your own group needs to seem better than other groups, so you try to engineer this by making other groups look worse in comparison. This in-group favouristim and discrimination stems from prejudice caused by grouping.

Reliability

Strength - Tajfel used the laboratory experiment method of research which had the advantage of enabling him to control the environment in terms of what the participants experienced including the information and instructions given to them and ensure that no other factors could influence their behaviour. Manipulation of the environment in this way enables cause and effect relationships to be indicated and the use of standardised procedures makes replications possible.

Generalisability

Weakness - The sample could be said to be biased as the participants were all male, of similar and from one particular school in one particular area of the country. It is difficult to see how these results could be generalised to other groups such as females, other age groups or people from other geographical areas. It could also be argued that these teenage boys were simply displaying the competitiveness typical of boys this age and not discrimination.

Validity

Strength and Weakness - The data generated was quantitative in that it involved calculating the number of participants who selected the different options in the matrices. This data allowed for comparisons to be made and statistical analysis to be carried out. However the study provided no qualitative data describing how the boys behaved or why they made the choices they did. So the study lacks validity.

Ecological validity

The study can be criticised in terms of its low ecological validity because of the unusual task that was performed in an artificial environment and it could be argued that it produced unnatural behaviour on the part of the participants. There is also a strong possibility that they were influenced by the demand characteristics of the situation and acted in the way that they thought was expected of them.

Reductionism

Another criticism is that Tajfel has reduced the complex human behaviour of discrimination to a very simple level, focusing just on minimal groups and performance of a simple experimental task.

Application

This study deals with a destructive and anti-social, but very common feature of society in examining the causes of prejudice and discrimination. As such it can be seen as a very useful piece of research that could be used to improve everyday life. However, the applications are restricted by the methodological limitations such as low ecological validity, and unrepresentative sample and the reductionist principles adopted.